Cross-cultural differences in visual object and background processing in the infant brain

Abstract

Human visual cognition differs profoundly between cultures. A key finding is that visual processing is tuned toward focal elements of a visual scene in Western cultures (US and Europe) and toward the background in Eastern cultures (Asia). Although some evidence for cultural differences exists for young children, to date, the ontogenetic origins of cultural differences in human visual cognition have not been unveiled. This study explores early cross-cultural differences in human visual processing, by tracking the neural signatures for object versus background elements of a visual scene in the electroencephalogram (EEG) of 12-month-old infants, in Vienna (Austria; a Western culture; n = 35) and Kyoto (Japan; an Eastern culture; n = 36). Specifically, we separated neural signatures by presenting object and background at different stimulation frequencies (5.67 and 8.5 Hz). Results show that human visual processing is different between cultures from early on. We found that infants from Vienna showed a higher object signal, in contrast to infants from Kyoto, who showed an accentuated background signal. This early emergence of cultural differences in human vision may be explained in part by early social experiences: In a separate interaction phase, mothers from Vienna pointed out object (versus background) elements more often than mothers from Kyoto. To conclude, with a cross-cultural developmental neuroscience approach, we reveal that cross-cultural differences in visual processing of object and background are already present in the first year after birth, which is much earlier than previously thought.

Keywords: cross-cultural comparison; frequency tagging; infant cognition; social learning; visual system development.

Fig. 1.

Experimental paradigm and grand mean neural response. (A) Twenty natural pictures with a clear object in front of a background were shown in a pre- and post-phase (3 s each) and during an interaction phase (6 s, each picture presented twice). In the interaction phase, the mother freely pointed out elements of the pictures to the infant. Before each image, a black screen was shown for EEG baseline recording (1 s) followed by an attention-getter, a white fixation dot with variable duration (0.5-1 s) accompanied by an infant-friendly sound. (B) By using frequency tagging, object and background were presented at different stimulation frequencies (5.67 or 8.5 Hz, counterbalanced), to elicit distinct neural signatures. (C) Frequency tagging led to a significant increase in the signal-to-noise ratio (SNR) compared to 1 at both stimulation frequencies (0-3000 ms, ***p <.001). (D) The SNR at the participants’ peak frequencies (ranges: 4-6 and 7-9 Hz; displayed up to 1.15 and 1.35 respectively) revealed the strongest signal at central, parietal, temporal, and occipital electrodes (black circles).

Fig. 2.

Topographies for (A) the object, the background, and (B) the object-background difference in the neural response of infants from Vienna and Kyoto. Values indicate the signal-to-noise ratio (SNR) at individualized frequencies, averaged across both phases (pre, post), for the whole time window of stimulus presentation (0-3000 ms). (C) The relative activity for the object versus the background (object score, for the electrodes marked in Fig. 1D), calculated as the SNR of the object, divided by the mean SNR of the object and background (whiskers indicate standard errors, the gray line at 1.00 corresponds to an equal signal strength for object and background), main effect Culture: ***p <.001.

Fig. 3.

Maternal pointing to the object versus the background. Violin plots indicate the proportion of points that mothers made to the object, and dots indicate individual participants. Mothers from Vienna pointed more often to the object than mothers from Kyoto, **p = .002.